US to make improvements in facial recognition software

The US federal government is closer to making a surveillance system that would pair computers with video cameras to scan crowds and automatically identify people by their faces.

The New York Times has got its paws on reports relating to the Department of Homeland Security’s crowd scanning project called the Biometric Optical Surveillance System [BOSS].

The system was not fully baked, but researchers say they are making significant advances.

The idea is to have a system that would help match faces in a crowd with names on a watch list. It would allow police to spot suspects at high-profile events like a presidential inaugural parade or find those who have escaped from prison.

So far technical specialists say crowd scanning is still too slow and unreliable, but the NSA is still keen on coming up with a workable plan in time for when computer processing power is strong enough.

BOSS began as an effort to help the military detect potential suicide bombers and other terrorists overseas at “outdoor polling places in Afghanistan and Iraq”. But in 2010, the Department of Homeland Security wanted it to be developed for use by the police in the United States.

During a recent test of the system, the department recommended against deploying it until more improvements could be made. However, at this point it is not clear when this may be done.

Researchers say they made progress, and independent specialists say it is virtually inevitable that someone will make the broader concept work as camera and compute power continue to improve.

The feeling is that it could be in place in five years centred around a two-year, $5.2 million federal contract given to Electronic Warfare Associates.

BOSS is handicapped by the fact that taking snaps of crowds from a distance is blighted with lighting problems and faces tend to be partly obscured.

Currently BOSS researchers are trying to overcome those challenges by generating far more information for computers to analyse.

BOSS consists of two towers bearing “robotic camera structures” with infrared and distance sensors. They take pictures of the same subject from slightly different angles.

A computer then processes the images into a “3D signature” built from data like the ratios between various points on someone’s face to be compared against database of faces.

A more recent test used 30 volunteers whose facial data would be mingled in a database among 1,000 mug shots. The agency set up six tests to determine the technology’s overall accuracy. The test worked out that the technology was not ready for police to buy.

Currently the belief is that the technology could be ready to deploy within five years.